1. Field of Invention
The present invention relates generally to improved methods of and apparatus and instruments for acquiring and analyzing information about the physical attributes of objects (such as reading bar code labels affixed to such objects), and digital image analysis.
2. Brief Description of the State of Knowledge in the Art
The use of image-based bar code symbol readers and scanners is well known in the field of auto-identification. Examples of image-based bar code symbol reading/scanning systems include, for example, hand-hand scanners, point-of-sale (POS) scanners, and industrial-type conveyor scanning systems.
Presently, most commercial image-based bar code symbol readers are constructed using charge-coupled device (CCD) image sensing/detecting technology. Unlike laser-based technology, CCD imaging technology has particular illumination requirements which differ from application to application.
For example, most prior art CCD-based hand-held bar code symbol readers utilize an array of light emitting diodes (LEDs) to flood the field of view of the imaging optics. In such systems, a large percentage of the output illumination from these LED sources is dispersed to regions other than the field of view of the imaging optics. Consequently, only a small percentage of the illumination is actually collected by the imaging optics of the system, and a large percentage of the illumination is wasted.
Examples of such prior art CCD-based hand-held bar code symbol readers are disclosed in U.S. Pat. Nos. Re. 36,528, 5,777,314, 5,756,981, 5,627,358, 5,484,994, 5,786,582, and 6,123,261 to Roustaei, each assigned to Symbol Technologies, Inc. and incorporated herein by reference in its entirety. In such prior art CCD-based hand-held scanners, an array of LEDs are mounted in a scanning head in front of a CCD image sensor that is provided with a cylindrical lens assembly. The LEDs are arranged at an angular orientation relative to a central axis passing through the scanning head so that a fan of light is emitted through the light transmission aperture thereof that expands with increasing distance away from the LEDs. The intended purpose of this LED illumination arrangement is to increase the “angular distance” and “depth of field” of such hand-held image-based bar code symbol readers. However, even with such improvements in LED illumination techniques, the working distance of such hand-held scanners can only be extended by using more LEDs within the scanning head of such scanners to produce greater illumination output therefrom, thereby increasing the cost, size and weight of such scanning devices.
Other CCD-based hand-held bar code symbol readers employing LED illumination have been proposed. For example, U.S. Pat. No. 5,192,856 to Schaham discloses a hand-held image scanner which uses a LED and beam forming optics (which include collimating and cylindrical lenses) to produce a beam of LED-based illumination for illuminating a bar code symbol on an object, and cylindrical optics mounted in front a linear CCD image detector for projecting a narrow a field of view about the illumination, thereby enabling collection and focusing of light reflected off the bar code symbol onto the linear CCD image detector.
CCD-based hand-held bar code symbol readers employing CCD image laser illumination have been proposed. For example, U.S. Pat. No. 4,963,756 to Quan et al discloses a hand-held image scanner using a laser source and Scheimpflug optics for focusing a planar laser illumination beam reflected off a bar code symbol onto a 2-D CCD image detector. U.S. Pat. No. 5,621,203 to Swartz et al discloses the use of a cylindrical lens to generate from a single laser diode an elongated beam of laser light. The fixed, static elongated beam is redirected by an oscillating mirror or lens such that it fans out an angle sufficient to illuminate a code pattern at a working distance and is swept in a direction transverse to the elongated dimension of the beam. A lens is mounted before a linear CCD image array, to receive diffused reflected laser light from the bar code symbol surface. And U.S. Pat. No. 5,988,506 to Schaham et al, herein incorporated by reference, discloses the use of a cylindrical lens to generate from a single visible laser diode (VLD) a narrow focused line of laser light which fans out an angle sufficient to fully illuminate a code pattern at a working distance. As disclosed, mirrors can be used to fold the laser illumination beam towards the code pattern to be illuminated in the working range of the system. Also, a horizontal linear lens array consisting of lenses is mounted before a linear,CCD image array, to receive diffused reflected laser light from the code symbol surface. Each single lens in the linear lens array forms its own image of the code line illuminated by the laser illumination beam. Also, subaperture diaphragms are required in the CCD array plane to (i) differentiate image fields, (ii) prevent diffused reflected laser light from passing through a lens and striking the image fields of neighboring lenses, and (iii) generate partially-overlapping fields of view from each of the neighboring elements in the lens array.
Most prior art CCD-based image scanners employed in conveyor-type package identification systems require high-pressure sodium, metal halide or halogen lamps and large, heavy and expensive parabolic or elliptical reflectors to produce sufficient light intensities to illuminate the large depth of field scanning fields supported by such industrial scanning systems. Even when the light from such lamps is collimated or focused using such reflectors, light strikes the target object other than where the imaging optics of the CCD-based camera are viewing. Since only a small fraction of the lamps output power is used to illuminate the CCD camera's field of view, the total output power of the lamps must be very high to obtain the illumination levels required along the field of view of the CCD camera. The balance of the output illumination power is simply wasted in the form of heat.
U.S. Provisional Application No. 60/190,273 entitled “Coplanar Camera” filed Mar. 17, 2000, by Chaleff et al., and published by WIPO on Sep. 27, 2001 as part of WIPO Publication No. WO 01/72028 A1, both being incorporated herein by reference, discloses a CCD camera system which uses an array of LEDs and a single apertured Fresnel-type cylindrical lens element to produce a planar beam of illumination for illuminating a bar code symbol on an object over a large depth of field, and a linear CCD image detector mounted behind the apertured Fresnel-type cylindrical lens element so as to provide the linear CCD image detector with a field of view that is arranged with the planar extent of planar beam of LED-based illumination.
While the prior art laser-illuminated CCD-based image capture systems discussed above avoid the use of LED illumination, they suffer from several significant shortcomings and drawbacks. For example, when detecting images of target objects illuminated by a coherent illumination source (e.g. a VLD), “speckle” (i.e. substrate or paper) noise is typically modulated onto the laser illumination beam during reflection/scattering, and ultimately speckle-noise patterns are produced at the CCD image detection array, severely reducing the signal-to-noise (SNR) ratio of the CCD camera system. Importantly, the prior art systems described above fail to provide any way of, or means for reducing speckle-noise patterns produced at its CCD image detector thereof, by its coherent laser illumination source.
In general, speckle-noise patterns are generated whenever the phase of the optical field is randomly modulated. The problem of speckle-noise patterns in laser scanning systems is mathematically analyzed in the twenty-five (25) slide show entitled “Speckle Noise and Laser Scanning Systems” by Sasa Kresic-Juric, Emanuel Marom and Leonard Bergstein, of Symbol Technologies, Holtsville, N.Y., published at http://www.ima.umn.edu/industrial/99-2000/kresic/sld001.htm, and incorporated herein by reference. Notably, Slide 11/25 of this WWW publication summaries two generally well known methods of reducing speckle-noise by superimposing statistically independent (time-varying) speckle-noise patterns: (1) using multiple laser beams to illuminate different regions of the speckle-noise scattering plane (i.e. object); or (2) using multiple laser beams with different wavelengths to illuminate the scattering plane. Also, the celebrated textbook by J. C. Dainty, et al, entitled “Laser Speckle and Related Phenomena” (Second edition), published by Springer-Verlag, 1994, incorporated herein by reference, describes a collection of techniques which have been developed by others over the years in effort to reduce speckle-noise. However, the prior art generally fails to disclose, teach or suggest how such prior art speckle-reduction techniques might be successfully practiced in a laser illuminated hand-held bar code symbol reader.
As described above, hand-held image-based bar code symbol readers may utilize a linear (1-D) imaging array or an area (2-D) imaging array to receive diffused reflected laser light from the bar code symbol surface. Advantageously, a linear imaging array provides a lower component cost. However, the use of the lower-cost linear imaging array in capturing images introduces problems. One problem is aspect ratio variations/distortions in the images captured by the linear imaging array. More specifically, in the event the scanning beam/imaging device is moved with respect to the bar code label/object to be scanned at a varying velocity during image capture operations, the image captured by the linear imaging array is distorted as shown in FIGS. 1A and 1B. Where the scanning beam/linear imaging array is moved with respect to the bar code label/object to be scanned at increasing velocity, the image is compressed as shown. On the other hand, where the scanning beam/linear imaging array is moved with respect to the bar code label/object to be scanned at decreasing velocity, the image is expanded as shown. If such distortion is significant, it can render the bar code symbol reader ineffective in many applications, such as reading dense 2-D bar code symbols. Another problem is jitter (motion of the object relative to device in a direction transverse to the intended swipe direction). Such transverse motion is referred to herein as “transverse jitter” or “horizontal jitter”. FIG. 15A illustrates the distortion causes by such horizontal jitter. If such horizontal jitter is significant, the resulting image distortions can degrade the quality of the images captured by the device and lead to errors/inaccuracies in subsequent image analysis operations (such as bar code symbol detections operations and/or OCR operations) performed on these images.
Similar problems occur in hand-held image capture devices, examples of which are described in U.S. Pat. Nos. 5,578,813; 6,222,174; and 6,300,645. Such hand-held image capture devices use a mechanical position transducer (e.g., roller or wheel) that operates similar to a computer mouse to output position information of the device relative to the object. Alternately, such hand-held image capture devices use one or more optical sensors that operate similar to an optical mouse (for example, as described in U.S. Pat. Nos. 4,631,400; 4,794,384; 5,729,008 and 6,256,016) to output position information of the device relative to the object. In such image capture devices, the position information is tagged to image data derived from an imaging array during image capture operations to generate a position-tagged data stream, and image stitching techniques are applied to the position-tagged data stream to bring multiple image swaths into registration to thereby form a single image.
Similar problems occur in astronomy, target tracking, velocimetry and airborne reconnaissance, examples of which are set forth in U.S. Pat. Nos. 4,162,509, 4,580,894; 5,020,903. Such systems use correlation of pixel data values derived from two spaced-apart linear imaging sensors to measure velocity.
Importantly, the prior art image-based techniques for velocity measurement are passive (relying on ambient light) or employ light sources that illuminate over a wide area. In such configurations, light is dispersed to regions other than the field of view of the imaging optics. Consequently, only a small percentage of the illumination is actually collected by the imaging optics of the system, and a large percentage of the illumination is wasted.
Thus, there is a great need in the art for improved image-based techniques/devices for velocity measurement that do not waste illumination (and thus require less power to provide such illumination).
Moreover, the prior art generally fails to disclose, teach or suggest image-based techniques/devices that measure velocity variations of the imaging device with respect to the object to be scanned and compensate for aspect ratio distortions that result from such velocity variations, which are suitable for use in a hand-held bar code symbol reader. Moreover, the prior art generally fails to disclose, teach or suggest image-based techniques/devices that perform velocity estimation/aspect ratio compensation in addition to image-based jitter estimation and compensation to compensate for both aspect ratio distortion and jitter distortion, which are suitable for use in a hand-held bar code symbol reader. In such an environment, these image-based techniques must operate at high speeds and operate in a resource constrained processing environment—constrained in memory, power utilization and weight—in order to provide cost-effective distortion-free real-time image acquisition and image-based bar code symbol reading suitable for many diverse applications, such as reading dense 2-D bar code symbols.
Thus, there is a great need in the art for improved image-based techniques/devices that provide velocity estimation/aspect ratio compensation and image-based jitter estimation and compensation to compensate for aspect ratio distortions and jitter distortions, which are suitable for use in a hand-held bar code symbol reader in order to provide cost-effective distortion-free real-time image acquisition and image-based bar code symbol reading suitable for many diverse applications, such as reading dense 2-D bar code symbols.